Citation Metrics Explained: h-index, Impact Factor, h5-index and the Rest
Open your Google Scholar profile and you see an h-index and an i10-index. Look at a journal’s homepage and it advertises an Impact Factor. Check a conference on Google Scholar Metrics and it has an h5-index. Each of these is a citation metric, each measures something slightly different, and the same researcher or venue can post very different numbers depending on which database you ask. This guide explains the metrics you are most likely to run into, what each one actually measures, and how to read them when you are deciding where to publish or which venues matter in your field.
The three things metrics measure
Citation metrics fall into three groups depending on what they are scoring: an individual researcher, a publication venue (a conference or journal), or a single paper. Confusion usually comes from comparing a metric in one group against a metric in another. An author h-index and a journal Impact Factor are not measuring the same thing and cannot be compared directly.
Author-level metrics
These describe the output of an individual researcher across their whole career.
h-index. The most widely cited author metric. Your h-index is the largest number h such that you have h papers with at least h citations each. An h-index of 20 means you have 20 papers that have each been cited at least 20 times. It rewards consistent impact rather than a single famous paper, which is why it has become the default shorthand for a research career. Its main weakness is that it never goes down and grows naturally with career length, so it favours senior researchers and is hard to compare across fields with different citation cultures.
i10-index. A Google Scholar invention: the number of your publications with at least 10 citations each. It is simpler than the h-index and only appears on Google Scholar profiles. It is a rough measure of how much of your output has gained traction, but it is rarely used in formal evaluation.
Total citations. The raw count of every citation to all of your work. Easy to understand and easy to inflate, since one highly cited paper or a long career can dominate the number. Useful as context alongside the h-index rather than on its own.
Venue-level metrics for conferences
This is the group that matters most when you are choosing where to submit, especially in computer science where conferences are the primary publication venue rather than journals.
h5-index. Google Scholar Metrics ranks publication venues using the h5-index, which is simply the h-index calculated over the articles a venue published in the last five complete years. A conference with an h5-index of 80 published at least 80 papers in the past five years that have each been cited at least 80 times. Because it uses a rolling five-year window, it reflects a venue’s recent influence rather than its historical reputation, and it is one of the better quick signals for how active and influential a conference is right now.
h5-median. Reported alongside the h5-index, this is the median number of citations for the papers that make up the h5-index. It tells you whether the citations are spread evenly across those top papers or concentrated in a handful. Two venues with the same h5-index can have very different h5-medians, and the higher median usually indicates a stronger, more consistent body of work.
For computer science specifically, the h5-index and a venue’s CORE ranking together give you a much fuller picture than either alone. CORE captures expert and community judgement of prestige, while the h5-index captures recent measured citation impact. A venue that is strong on both is a safe target. You can see the CORE rank on every conference page on WorkWander, and cross-reference the h5-index on Google Scholar Metrics.
Journal-level metrics
If your work fits a journal rather than a conference, or you work in a field where journals dominate, these are the numbers you will see.
Impact Factor (JIF). The oldest and most famous journal metric, published by Clarivate in the Journal Citation Reports. A journal’s Impact Factor for a given year is the number of citations that year to articles it published in the previous two years, divided by the number of citable articles in those two years. It is based on the curated Web of Science database. It is influential but heavily criticised: a two-year window is short for slow-moving fields, review articles inflate it, and a single blockbuster paper can swing a small journal’s number significantly.
CiteScore. Elsevier’s answer to the Impact Factor, calculated from the Scopus database. It uses a four-year citation window instead of two, which makes it more stable and more generous in coverage. Because it draws on a different database than the Impact Factor, the same journal will usually show different CiteScore and Impact Factor values.
SJR (SCImago Journal Rank). Also based on Scopus, but instead of counting all citations equally it weights them by the prestige of the citing journal, in the same spirit as Google’s PageRank. A citation from a highly regarded journal counts for more than one from an obscure one. This tries to measure influence rather than raw volume.
SNIP (Source Normalized Impact per Paper). Another Scopus metric, designed to correct for the fact that some fields simply cite more than others. It normalises a journal’s citations against the citation potential of its field, which makes cross-field comparison more meaningful than a raw Impact Factor allows.
Eigenfactor. Based on Web of Science, this scores a journal by its total influence on the literature, weighting citations by the influence of the citing journals and accounting for journal size differently than the Impact Factor. It is reported alongside the Impact Factor in the Journal Citation Reports.
Why the same metric gives different numbers
The single most confusing thing about citation metrics is that your h-index on Google Scholar is almost always higher than your h-index on Scopus or Web of Science. This is not an error. Each metric is computed over a different underlying database, and the databases cover different things.
Google Scholar has the broadest coverage. It indexes conference papers, preprints on arXiv, theses, technical reports, and books, and it counts citations from all of them. This makes it the most generous source and the best for fields like computer science where conferences and preprints carry real weight. The trade-off is less quality control, occasional duplicate or phantom records, and no manual curation.
Scopus, run by Elsevier, is a large curated database. Its coverage is narrower than Google Scholar but broader than Web of Science, and it is the source for CiteScore, SJR, SNIP, and field-weighted metrics.
Web of Science, run by Clarivate, is the oldest and most selective database. It indexes a smaller, more tightly curated set of journals and conferences, so citation counts drawn from it are typically the lowest of the three. It is the source for the Impact Factor and the Eigenfactor.
Semantic Scholar, a free database from the Allen Institute for AI, is strong in computer science and biomedicine and adds AI-driven features like flagging highly influential citations rather than treating every citation equally.
Because of these differences, never compare a number from one source against a number from another. A Google Scholar h-index of 30 and a Web of Science h-index of 22 can belong to the same researcher.
Field-Weighted Citation Impact (FWCI), a Scopus metric, is worth knowing as the cleanest attempt to make a single paper’s impact comparable across fields. It is the ratio of citations a paper actually received to the average number received by similar papers of the same type, field, and age. A value of 1.0 is exactly world average, and 2.0 is twice the world average, regardless of whether the field cites heavily or sparingly.
What the metrics miss
Every metric here counts citations, and citations are an imperfect proxy for quality. They accumulate slowly, so recent important work looks weak for years. They vary enormously by field, so a strong number in mathematics looks small next to an average number in biomedicine. They can be gamed through self-citation and citation rings. And they say nothing about whether a citation was approving, critical, or a passing mention.
They also miss entire categories of impact. Released software, datasets, reproducible artifacts, influence on industry practice, and teaching contributions rarely show up in any citation count, even when they are the most valuable thing a researcher produces.
How to actually use them
Use citation metrics as one input, not a verdict. When you are choosing where to submit, combine a venue’s h5-index with its CORE ranking and, most importantly, with where your specific research community actually publishes and gathers. When you are reading your own profile, treat the h-index as a rough career marker rather than a target to optimise. When you are comparing two researchers or two venues, make sure you are comparing the same metric from the same database, and weight the comparison by the norms of the field.
The numbers are useful precisely because they are quick. They are dangerous when they become the goal instead of the signal. To see CORE rankings, dates, and venue details for the conferences in your field, browse the full conference calendar on WorkWander.